2018-01-29 13:11:24 +00:00
|
|
|
/* SPDX-License-Identifier: BSD-3-Clause
|
|
|
|
* Copyright 2016 6WIND S.A.
|
2016-10-03 08:38:46 +00:00
|
|
|
*/
|
|
|
|
|
|
|
|
#ifndef _RTE_NET_PTYPE_H_
|
|
|
|
#define _RTE_NET_PTYPE_H_
|
|
|
|
|
|
|
|
#ifdef __cplusplus
|
|
|
|
extern "C" {
|
|
|
|
#endif
|
|
|
|
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
#include <rte_ip.h>
|
|
|
|
#include <rte_udp.h>
|
|
|
|
#include <rte_tcp.h>
|
|
|
|
#include <rte_sctp.h>
|
|
|
|
|
2016-10-03 08:38:46 +00:00
|
|
|
/**
|
|
|
|
* Structure containing header lengths associated to a packet, filled
|
|
|
|
* by rte_net_get_ptype().
|
|
|
|
*/
|
|
|
|
struct rte_net_hdr_lens {
|
|
|
|
uint8_t l2_len;
|
|
|
|
uint8_t l3_len;
|
|
|
|
uint8_t l4_len;
|
|
|
|
uint8_t tunnel_len;
|
|
|
|
uint8_t inner_l2_len;
|
|
|
|
uint8_t inner_l3_len;
|
|
|
|
uint8_t inner_l4_len;
|
|
|
|
};
|
|
|
|
|
2018-03-28 15:43:42 +00:00
|
|
|
/**
|
|
|
|
* Skip IPv6 header extensions.
|
|
|
|
*
|
|
|
|
* This function skips all IPv6 extensions, returning size of
|
|
|
|
* complete header including options and final protocol value.
|
|
|
|
*
|
|
|
|
* @warning
|
|
|
|
* @b EXPERIMENTAL: this API may change without prior notice
|
|
|
|
*
|
|
|
|
* @param proto
|
|
|
|
* Protocol field of IPv6 header.
|
|
|
|
* @param m
|
|
|
|
* The packet mbuf to be parsed.
|
|
|
|
* @param off
|
|
|
|
* On input, must contain the offset to the first byte following
|
|
|
|
* IPv6 header, on output, contains offset to the first byte
|
|
|
|
* of next layer (after any IPv6 extension header)
|
|
|
|
* @param frag
|
|
|
|
* Contains 1 in output if packet is an IPv6 fragment.
|
|
|
|
* @return
|
|
|
|
* Protocol that follows IPv6 header.
|
|
|
|
* -1 if an error occurs during mbuf parsing.
|
|
|
|
*/
|
2019-06-29 11:58:53 +00:00
|
|
|
__rte_experimental
|
|
|
|
int
|
2018-03-28 15:43:42 +00:00
|
|
|
rte_net_skip_ip6_ext(uint16_t proto, const struct rte_mbuf *m, uint32_t *off,
|
|
|
|
int *frag);
|
|
|
|
|
2016-10-03 08:38:46 +00:00
|
|
|
/**
|
|
|
|
* Parse an Ethernet packet to get its packet type.
|
|
|
|
*
|
|
|
|
* This function parses the network headers in mbuf data and return its
|
|
|
|
* packet type.
|
|
|
|
*
|
|
|
|
* If it is provided by the user, it also fills a rte_net_hdr_lens
|
|
|
|
* structure that contains the lengths of the parsed network
|
|
|
|
* headers. Each length field is valid only if the associated packet
|
|
|
|
* type is set. For instance, hdr_lens->l2_len is valid only if
|
|
|
|
* (retval & RTE_PTYPE_L2_MASK) != RTE_PTYPE_UNKNOWN.
|
|
|
|
*
|
|
|
|
* Supported packet types are:
|
2016-10-03 08:38:48 +00:00
|
|
|
* L2: Ether, Vlan, QinQ
|
2016-10-03 08:38:46 +00:00
|
|
|
* L3: IPv4, IPv6
|
|
|
|
* L4: TCP, UDP, SCTP
|
2016-10-03 08:38:52 +00:00
|
|
|
* Tunnels: IPv4, IPv6, Gre, Nvgre
|
2016-10-03 08:38:46 +00:00
|
|
|
*
|
|
|
|
* @param m
|
|
|
|
* The packet mbuf to be parsed.
|
|
|
|
* @param hdr_lens
|
|
|
|
* A pointer to a structure where the header lengths will be returned,
|
|
|
|
* or NULL.
|
2016-10-03 08:38:53 +00:00
|
|
|
* @param layers
|
|
|
|
* List of layers to parse. The function will stop at the first
|
|
|
|
* empty layer. Examples:
|
|
|
|
* - To parse all known layers, use RTE_PTYPE_ALL_MASK.
|
|
|
|
* - To parse only L2 and L3, use RTE_PTYPE_L2_MASK | RTE_PTYPE_L3_MASK
|
2016-10-03 08:38:46 +00:00
|
|
|
* @return
|
|
|
|
* The packet type of the packet.
|
|
|
|
*/
|
|
|
|
uint32_t rte_net_get_ptype(const struct rte_mbuf *m,
|
2016-10-03 08:38:53 +00:00
|
|
|
struct rte_net_hdr_lens *hdr_lens, uint32_t layers);
|
2016-10-03 08:38:46 +00:00
|
|
|
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
/**
|
|
|
|
* Prepare pseudo header checksum
|
|
|
|
*
|
|
|
|
* This function prepares pseudo header checksum for TSO and non-TSO tcp/udp in
|
|
|
|
* provided mbufs packet data and based on the requested offload flags.
|
|
|
|
*
|
|
|
|
* - for non-TSO tcp/udp packets full pseudo-header checksum is counted and set
|
|
|
|
* in packet data,
|
|
|
|
* - for TSO the IP payload length is not included in pseudo header.
|
|
|
|
*
|
|
|
|
* This function expects that used headers are in the first data segment of
|
|
|
|
* mbuf, are not fragmented and can be safely modified.
|
|
|
|
*
|
|
|
|
* @param m
|
|
|
|
* The packet mbuf to be fixed.
|
|
|
|
* @param ol_flags
|
|
|
|
* TX offloads flags to use with this packet.
|
|
|
|
* @return
|
|
|
|
* 0 if checksum is initialized properly
|
|
|
|
*/
|
|
|
|
static inline int
|
|
|
|
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
|
|
|
|
{
|
2019-06-28 03:13:09 +00:00
|
|
|
/* Initialise ipv4_hdr to avoid false positive compiler warnings. */
|
|
|
|
struct rte_ipv4_hdr *ipv4_hdr = NULL;
|
2019-05-21 16:13:10 +00:00
|
|
|
struct rte_ipv6_hdr *ipv6_hdr;
|
2019-05-21 16:13:13 +00:00
|
|
|
struct rte_tcp_hdr *tcp_hdr;
|
2019-05-21 16:13:14 +00:00
|
|
|
struct rte_udp_hdr *udp_hdr;
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
uint64_t inner_l3_offset = m->l2_len;
|
|
|
|
|
2019-03-29 13:42:14 +00:00
|
|
|
#ifdef RTE_LIBRTE_ETHDEV_DEBUG
|
|
|
|
/*
|
|
|
|
* Does packet set any of available offloads?
|
|
|
|
* Mainly it is required to avoid fragmented headers check if
|
|
|
|
* no offloads are requested.
|
|
|
|
*/
|
|
|
|
if (!(ol_flags & PKT_TX_OFFLOAD_MASK))
|
|
|
|
return 0;
|
|
|
|
#endif
|
|
|
|
|
2019-05-24 14:57:14 +00:00
|
|
|
if (ol_flags & (PKT_TX_OUTER_IPV4 | PKT_TX_OUTER_IPV6))
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
inner_l3_offset += m->outer_l2_len + m->outer_l3_len;
|
|
|
|
|
2019-03-29 13:42:14 +00:00
|
|
|
#ifdef RTE_LIBRTE_ETHDEV_DEBUG
|
|
|
|
/*
|
|
|
|
* Check if headers are fragmented.
|
|
|
|
* The check could be less strict depending on which offloads are
|
|
|
|
* requested and headers to be used, but let's keep it simple.
|
|
|
|
*/
|
|
|
|
if (unlikely(rte_pktmbuf_data_len(m) <
|
|
|
|
inner_l3_offset + m->l3_len + m->l4_len))
|
|
|
|
return -ENOTSUP;
|
|
|
|
#endif
|
|
|
|
|
2018-09-19 14:42:08 +00:00
|
|
|
if (ol_flags & PKT_TX_IPV4) {
|
2019-05-21 16:13:10 +00:00
|
|
|
ipv4_hdr = rte_pktmbuf_mtod_offset(m, struct rte_ipv4_hdr *,
|
2018-09-19 14:42:08 +00:00
|
|
|
inner_l3_offset);
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
|
2018-09-19 14:42:08 +00:00
|
|
|
if (ol_flags & PKT_TX_IP_CKSUM)
|
|
|
|
ipv4_hdr->hdr_checksum = 0;
|
|
|
|
}
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
|
2019-06-28 03:13:09 +00:00
|
|
|
if ((ol_flags & PKT_TX_L4_MASK) == PKT_TX_UDP_CKSUM) {
|
2018-09-19 14:42:08 +00:00
|
|
|
if (ol_flags & PKT_TX_IPV4) {
|
2019-05-21 16:13:14 +00:00
|
|
|
udp_hdr = (struct rte_udp_hdr *)((char *)ipv4_hdr +
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
m->l3_len);
|
|
|
|
udp_hdr->dgram_cksum = rte_ipv4_phdr_cksum(ipv4_hdr,
|
|
|
|
ol_flags);
|
|
|
|
} else {
|
2019-05-21 16:13:10 +00:00
|
|
|
ipv6_hdr = rte_pktmbuf_mtod_offset(m,
|
|
|
|
struct rte_ipv6_hdr *, inner_l3_offset);
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
/* non-TSO udp */
|
2019-05-21 16:13:14 +00:00
|
|
|
udp_hdr = rte_pktmbuf_mtod_offset(m,
|
|
|
|
struct rte_udp_hdr *,
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
inner_l3_offset + m->l3_len);
|
|
|
|
udp_hdr->dgram_cksum = rte_ipv6_phdr_cksum(ipv6_hdr,
|
|
|
|
ol_flags);
|
|
|
|
}
|
2019-06-28 03:13:09 +00:00
|
|
|
} else if ((ol_flags & PKT_TX_L4_MASK) == PKT_TX_TCP_CKSUM ||
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
(ol_flags & PKT_TX_TCP_SEG)) {
|
|
|
|
if (ol_flags & PKT_TX_IPV4) {
|
|
|
|
/* non-TSO tcp or TSO */
|
2019-05-21 16:13:13 +00:00
|
|
|
tcp_hdr = (struct rte_tcp_hdr *)((char *)ipv4_hdr +
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
m->l3_len);
|
|
|
|
tcp_hdr->cksum = rte_ipv4_phdr_cksum(ipv4_hdr,
|
|
|
|
ol_flags);
|
|
|
|
} else {
|
2019-05-21 16:13:10 +00:00
|
|
|
ipv6_hdr = rte_pktmbuf_mtod_offset(m,
|
|
|
|
struct rte_ipv6_hdr *, inner_l3_offset);
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
/* non-TSO tcp or TSO */
|
2019-05-21 16:13:13 +00:00
|
|
|
tcp_hdr = rte_pktmbuf_mtod_offset(m,
|
|
|
|
struct rte_tcp_hdr *,
|
ethdev: add Tx preparation
Added API for `rte_eth_tx_prepare`
uint16_t rte_eth_tx_prepare(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
Added fields to the `struct rte_eth_desc_lim`:
uint16_t nb_seg_max;
/**< Max number of segments per whole packet. */
uint16_t nb_mtu_seg_max;
/**< Max number of segments per one MTU */
These fields can be used to create valid packets according to the
following rules:
* For non-TSO packet, a single transmit packet may span up to
"nb_mtu_seg_max" buffers.
* For TSO packet the total number of data descriptors is "nb_seg_max",
and each segment within the TSO may span up to "nb_mtu_seg_max".
Added functions:
int
rte_validate_tx_offload(struct rte_mbuf *m)
to validate general requirements for tx offload set in mbuf of packet
such a flag completness. In current implementation this function is
called optionaly when RTE_LIBRTE_ETHDEV_DEBUG is enabled.
int rte_net_intel_cksum_prepare(struct rte_mbuf *m)
to prepare pseudo header checksum for TSO and non-TSO tcp/udp packets
before hardware tx checksum offload.
- for non-TSO tcp/udp packets full pseudo-header checksum is
counted and set.
- for TSO the IP payload length is not included.
int
rte_net_intel_cksum_flags_prepare(struct rte_mbuf *m, uint64_t ol_flags)
this function uses same logic as rte_net_intel_cksum_prepare, but
allows application to choose which offloads should be taken into
account, if full preparation is not required.
PERFORMANCE TESTS
-----------------
This feature was tested with modified csum engine from test-pmd.
The packet checksum preparation was moved from application to Tx
preparation step placed before burst.
We may expect some overhead costs caused by:
1) using additional callback before burst,
2) rescanning burst,
3) additional condition checking (packet validation),
4) worse optimization (e.g. packet data access, etc.)
We tested it using ixgbe Tx preparation implementation with some parts
disabled to have comparable information about the impact of different
parts of implementation.
IMPACT:
1) For unimplemented Tx preparation callback the performance impact is
negligible,
2) For packet condition check without checksum modifications (nb_segs,
available offloads, etc.) is 14626628/14252168 (~2.62% drop),
3) Full support in ixgbe driver (point 2 + packet checksum
initialization) is 14060924/13588094 (~3.48% drop)
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2016-12-23 18:40:47 +00:00
|
|
|
inner_l3_offset + m->l3_len);
|
|
|
|
tcp_hdr->cksum = rte_ipv6_phdr_cksum(ipv6_hdr,
|
|
|
|
ol_flags);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Prepare pseudo header checksum
|
|
|
|
*
|
|
|
|
* This function prepares pseudo header checksum for TSO and non-TSO tcp/udp in
|
|
|
|
* provided mbufs packet data.
|
|
|
|
*
|
|
|
|
* - for non-TSO tcp/udp packets full pseudo-header checksum is counted and set
|
|
|
|
* in packet data,
|
|
|
|
* - for TSO the IP payload length is not included in pseudo header.
|
|
|
|
*
|
|
|
|
* This function expects that used headers are in the first data segment of
|
|
|
|
* mbuf, are not fragmented and can be safely modified.
|
|
|
|
*
|
|
|
|
* @param m
|
|
|
|
* The packet mbuf to be fixed.
|
|
|
|
* @return
|
|
|
|
* 0 if checksum is initialized properly
|
|
|
|
*/
|
|
|
|
static inline int
|
|
|
|
rte_net_intel_cksum_prepare(struct rte_mbuf *m)
|
|
|
|
{
|
|
|
|
return rte_net_intel_cksum_flags_prepare(m, m->ol_flags);
|
|
|
|
}
|
|
|
|
|
2016-10-03 08:38:46 +00:00
|
|
|
#ifdef __cplusplus
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
|
|
#endif /* _RTE_NET_PTYPE_H_ */
|